- September 14, 2025
- By Maryland Today Staff
Whether it’s for help writing formal invitations, making a vacation itinerary or getting tips on spark plug replacement, most of us are finding more and more ways to use the sometimes surprising capabilities of artificial intelligence chatbots. A University of Maryland-designed chatbot has also popped up as part of a pilot program to provide virtual study assistants (VSAs) that deliver course-specific help in a few dozen classes over the last year.
The limited program put the university at the cutting edge of integrating AI into the classroom in a supervised manner, and now, a group of experts in education and computing has launched an innovative research project to evaluate the AIs’ in-class performance—and they’re looking for Fall 2025 instructors to help.
[Watch a video that explains how instructors engage with VSAs.]
“The anecdotal reports have been pretty good; now we want to quantify how well they really work and find the best ways to use them,” said Axel Persaud, who oversaw the pilot program as assistant vice president of enterprise engineering for the division of information technology (DIT).
The study will be conducted by associate professors of education Tracy Sweet of the Department of Human Development and Quantitative Methodology and Jing Liu of the Department of Teaching and Learning, Policy and Leadership (also the director of the Center for Education Data Science and Innovation), along with Megan C. Masters, DIT interim assistant vice president for academic technology and innovation.
To carry out the study, the researchers plan to observe sections of large courses with at least two instructors, in which some course sections will be randomly assigned to use the chatbot study assistant while others will not. They’ll track initial metrics like comparative performance in the course itself, but over the longer term, the study plans to monitor things such as long-term learning performance and how using a chatbot study assistant impacts students’ likelihood of continuing in their major or switching to another.
Very little research exists on how AI use affects student performance, and the new UMD study—likely the only one of its type—puts the university at the forefront of knowledge-gathering on the topic, the researchers said.
Instructors interested in participating in the study can fill out a form or contact Tracy Sweet at tsweet@umd.edu.